Golang httpRouter returns 与函数切片一起使用时的最后响应

Question

我正在尝试为 httprouter 包实现 expressjs 类似的功能。我创建了一个结构 type mounter

type Mounter struct {
    BasePath string
    Routes   []*Route
}

和一个 Route 结构，它表示子路由

type Route struct {
    Path   string
    Method string
    Func   Handle
}

type Handle func(http.ResponseWriter, *http.Request, Params)

type Params interface{}

我有一个 NewRoutes 函数，这是我想从 expressjs 新路由移植的主要功能，与 express.Router

做同样的事情

func NewRoutes(base string) (mounter *Mounter) {
    mounter = &Mounter{
        BasePath: base,
    }
    return
}

我在 *Mounter

下有 get post put delete 方法

//GET request handler
func (mounter *Mounter) GET(path string, Func Handle) {
    mounter.Routes = append(mounter.Routes, &Route{path, "get", Func})
}

//POST request handler
func (mounter *Mounter) POST(path string, Func Handle) {
    mounter.Routes = append(mounter.Routes, &Route{path, "post", Func})
}

//PUT request handler
func (mounter *Mounter) PUT(path string, Func Handle) {
    mounter.Routes = append(mounter.Routes, &Route{path, "put", Func})
}

//DELETE request handler
func (mounter *Mounter) DELETE(path string, Func Handle) {
    mounter.Routes = append(mounter.Routes, &Route{path, "delete", Func})
}

最后我有了一个将路由器安装到实际路由器的 Mount 方法

func (mounter *Mounter) Mount(router *rtr.Router) {
    mounter.BasePath = strings.TrimSuffix(mounter.BasePath, "/")
    for _, route := range mounter.Routes {
        path := route.Path
        if !strings.HasSuffix(path, "/") {
            path += "/"
        }
        path = mounter.BasePath + path
        switch strings.ToLower(route.Method) {
        case "get":
            router.GET(path, func(res http.ResponseWriter, req *http.Request, params rtr.Params) {
                route.Func(res, req, params)
            })
        case "post":
            router.POST(path, func(res http.ResponseWriter, req *http.Request, params rtr.Params) {
                route.Func(res, req, params)
            })
        case "delete":
            router.DELETE(path, func(res http.ResponseWriter, req *http.Request, params rtr.Params) {
                route.Func(res, req, params)
            })
        case "put":
            router.PUT(path, func(res http.ResponseWriter, req *http.Request, params rtr.Params) {
                route.Func(res, req, params)
            })
        }
    }
}

一切都很好，如果我尝试向 get 端点发送 post 请求，它会给出一个很好的 404，但唯一的问题是它总是响应最后一个处理程序添加成员而不考虑子路径所以

package api
var ApiRouter = express.NewRoutes("/api/")

func init() {
    ApiRouter.GET("/", func(res http.ResponseWriter, req *http.Request, _ express.Params) {
        fmt.Fprintln(res, "testget/")
    })
    ApiRouter.GET("/pt", func(res http.ResponseWriter, req *http.Request, _ express.Params) {
        fmt.Fprintln(res, "pt")
    })
    ApiRouter.POST("/test", func(res http.ResponseWriter, req *http.Request, _ express.Params) {
        fmt.Fprintln(res, "test/post")
    })
}

package main
func main() {
    router := express.New()
    api.ApiRouter.Mount(router)
    for _, route := range api.ApiRouter.Routes {
        fmt.Println(*route)
    }
    router.ServeFiles("/public/*filepath", http.Dir("./public/"))
    http.ListenAndServe(":1024", router)
}

将始终响应 test/post 并且我在上面出于测试目的所做的 range 的输出是那么你知道为什么它总是使用相同的函数来响应但完美地识别路径吗？

Answer 1

问题出在Mount方法上，即所谓的Iteration Variables and Closures。声明一个用于捕获 route 的新变量，例如

thisRoute := route //capture iteration variable
switch strings.ToLower(route.Method) {
case "get":
    router.GET(path, func(res http.ResponseWriter, req *http.Request, params rtr.Params) {
        thisRoute.Func(res, req, params)
    })
case "post":
    router.POST(path, func(res http.ResponseWriter, req *http.Request, params rtr.Params) {
        thisRoute.Func(res, req, params)
    })
//...
}

Golang httpRouter returns 与函数切片一起使用时的最后响应

Golang httpRouter returns the last response when used with the slice of functions

router

function-pointers

go

slice

express